[sdiy] 4069 CMOS VCO Trouble-Shooting

[René Schmitz]

Hi Michael, Scott and all,

Scott Gravenhorst wrote:
> "Michael David Brown" <1browm13 at solent.ac.uk> wrote:
>>I had this circuit generating the sawtooth output with freq' 
>>controlled by 22k Lin pot, 
> 
> 
> This should work as a test, but note that in practice, if you want this to scale
> correctly, the 22K pot should be replaced with a voltage controlled current source.

I think this is a misunderstanding, the 22k is the value I used for the 
normal tune pot, and I presume that is what Michael meant.

But it could be worthwhile to open the connection from the collector of 
T2 to the integrator gate. And tie a 1 Meg resistor from that integrator 
gates input to GND, to see if the actual core is working when it has 
some current to integrate.

If this turns out to be not the case, maybe if you measure the voltages, 
at the inputs and the outputs of the integrator and schmitt trigger 
gates, one could draw a conclusion as to where to look next.

Michael, don't give up. Electronics is not just about assembly and 
soldering. It is mostly about being able to troubleshoot the thing you 
have created.

> As far as layout goes, I did discover a sensitivity to which inverters were used
> for what purpose - and this was verified by René.  Note that it may affect only one
> manufacturer's die.  My problem was not the VCO itself, that always worked, rather
> I had problems with the PWM portion (noise burst instead of a rectangular edge). 
> After using the pin details I got from René, it worked perfectly.  You can see my
> pin details here:
> 
> http://home1.gte.net/res0658s/fatman/VCO_pwm_tri_suboctave.html

IIRC I could replicate that problem here. The problem occured because of 
stray coupling between a high impedance input and an adjacent output. 
Using the gates in their "natural" order, i.e. letting the signal flow 
from left to right through the gates should prevent this. Hence the note 
best to stick to the layout I gave on my page.

> Other than that, it ought to work.

I think so too. Especially because there are reports from many people 
who have sucessfully built it.

Cheers,
  René

-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159